HAL Id: hal-01137000
https://hal.archives-ouvertes.fr/hal-01137000
Submitted on 30 Mar 2015
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of
sci-entific research documents, whether they are
pub-lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Effects of endosulfan on hepatoma cell adhesion:
Epithelial-mesenchymal transition and anoikis resistance
Ludovic Peyre, Nathalie Zucchini-Pascal, Georges de Sousa, Roger Rahmani
To cite this version:
Ludovic Peyre, Nathalie Zucchini-Pascal, Georges de Sousa, Roger Rahmani. Effects of endosulfan
on hepatoma cell adhesion: Epithelial-mesenchymal transition and anoikis resistance. Toxicology,
Elsevier, 2012, 300 (1-2), pp.19-30. �10.1016/j.tox.2012.05.008�. �hal-01137000�
Contents
lists
available
at
SciVerse
ScienceDirect
Toxicology
j o u
r n
a l
h o m
e p a g e :
w w w . e l s e v i e r . c o m / l o c a t e / t o x i c o l
Effects
of
endosulfan
on
hepatoma
cell
adhesion:
Epithelial–mesenchymal
transition
and
anoikis
resistance
Ludovic
Peyre,
Nathalie
Zucchini-Pascal
∗
,
Georges
de
Sousa,
Roger
Rahmani
∗
LaboratoiredeToxicologieCellulaireetMoléculairedesXénobiotiques,INRA,UMR1331TOXALIMSophiaAntipolis,06903SophiaAntipolis,France
a
r
t
i
c
l
e
i
n
f
o
Articlehistory:Received28March2012
Receivedinrevisedform10May2012 Accepted12May2012
Available online 5 June 2012
Keywords: Endosulfan Anoikis EMT Hepatocarcinoma HepG2
a
b
s
t
r
a
c
t
EndosulfanisanorganochlorinepesticidecommonlyusedinagricultureyetclassifiedbytheStockholm Conventionin2011asapersistentorganicpollutant(POP).Itspotentialtoxicitymakesitscontinueduse amajorpublichealthconcern.Despitestudiesinlaboratoryanimals,themolecularmechanisms under-lyingthecarcinogeniceffectsofendosulfaninhumanliverremainpoorlyunderstood.Inthisstudy,we investigatedthephenotypicaleffectsofendosulfanonHepG2livercells.First,wefoundthatendosulfan disruptedtheanoikisprocess.Indeed,cellsexposedtoendosulfanwereinitiallysensitizedtoanoikisand thereafterrecoveredtheirresistancetothisprocess.Thisphenomenonoccurredinparalleltothe induc-tionoftheepithelialtomesenchymal(EMT)process,asdemonstratedby:(1)reorganizationoftheactin cytoskeletontogetherwithactivationoftheFAKsignalingpathway;(2)repressionofE-cadherin expres-sion;(3)inductionofSnailandSlug;(4)activationoftheWNT/-cateninpathway;and(5)inductionand reorganizationofmesenchymalmarkers(S100a4,vimentin,fibronectin,MMP-7).Secondly,despitethe acquisitionofmesenchymalcharacteristics,HepG2cellsexposedtoendosulfanfailedtomigrate.This incapacitytoacquireamotilephenotypecouldbeattributedtoadisruptionoftheinteractionbetween theECMandthecells.Takentogether,theseresultsindicatethatendosulfanprofoundlyaltersthe pheno-typeoflivercellsbyinducingcelldetachmentandpartialEMTaswellasdisruptingtheanoikisprocess. Alltheseeventsaccount,atleastinpart,forthecarcinogenicpotentialofendosulfaninliver.
© 2012 Elsevier Ireland Ltd. All rights reserved.
1.
Introduction
Since
the
1950s,
organochlorine
pesticides
(OCPs)
such
as
the
chlorinated
cyclodiene
endosulfan
have
been
employed
exten-sively
in
agriculture,
viticulture,
horticulture,
domestic
and
public
health.
Endosulfan
is
sold
under
different
trade
names
(Thiodan
®,
Rocky
®,
Thionex
®.
.
.)
and
corresponds
to
a
mixture
of
70%
endosulfan-
␣
and
30%
endosulfan-
.
While
it
does
have
beneficial
effects
on
crops,
it
also
represents
a
potential
source
of
environ-mental
contamination
and
risk
to
public
health
(
Silva
and
Gammon,
Abbreviations: Ab,antibody;DAPI,4,6-di-amidino-2-phenylindole;DMSO,
dimethylsulfoxide;ECM,extracellularmatrix;EMT,epithelialtomesenchymal tran-sition;FAK,focaladhesionkinase;FBS,foetalbovineserum;HCC,hepatocellular carcinoma;LEF1/TCF-1,lymphoidenhancerfactor1/Tcellfactor1;MMP, met-alloprotease;MTT,3-(4,5-dimethylthiazol-2-Yl)-2,5-diphenyltetrazoliumbromide; Poly-HEMA,poly2-hydroxyethylmethacrylate;POP,persistentorganicpollutant; ROCK,Rho-associatedproteinkinase;WNT,winglessintegrationsite;XIAP,X-linked inhibitorofapoptosisprotein.
∗ Correspondingauthorsat:LaboratoiredeToxicologieCellulaireetMoléculaire desXénobiotiques,INRA,UMR1331,400routedesChappes,BP167,06903Sophia Antipolis,France.Tel.:+33492386548;fax:+33492386401.
E-mailaddresses:zucchini@sophia.inra.fr(N.Zucchini-Pascal), rahmani@sophia.inra.fr(R.Rahmani).
2009
).
Indeed,
endosulfan
is
a
persistent
organic
pollutant
(POP)
that
can
build
up
in
the
environment
and
bioaccumulate
through
the
food
chain.
Its
persistent
and
bioaccumulative
properties
enable
it
to
become
stored
in
fats
where
it
persists
for
a
long
time.
Endo-sulfan
is
readily
absorbed
by
humans
via
the
stomach,
the
skin,
the
lungs,
and
the
placenta
(
Lopez-Espinosa
et
al.,
2008
).
This
OCP
is
one
of
the
most
toxic
pesticides
currently
on
the
market
and
is
responsible
for
a
large
number
of
poisoning-related
deaths
(
Moses
and
Peter,
2010
).
In
animals,
it
is
toxic
to
the
liver,
kidney,
nervous
system,
blood
system,
immune
system
and
reproductive
organs
(
Hashizume
et
al.,
2010;
Karatas
et
al.,
2006;
Ahmed
et
al.,
2011;
Merhi
et
al.,
2010;
Briz
et
al.,
2011;
Chan
et
al.,
2006;
Choudhary
et
al.,
2003;
Aggarwal
et
al.,
2008
).
Due
to
the
risk
of
adverse
effects
on
human
health,
the
World
Health
Organization
(WHO)
classi-fied
endosulfan
as
a
moderately
hazardous
Class
2
pesticide
while
the
Environmental
Protection
Agency
(EPA)
considered
it
as
“highly
acutely
toxic”
(Category
I).
Despite
then
being
banned
in
more
than
62
countries
and
by
the
Stockholm
Convention
in
2011,
it
is
still
used
extensively
(
Weber
et
al.,
2010
).
An
interdiction
for
use
will
be
effective
in
2017.
While
the
potential
carcinogenic
effects
of
endosulfan
remain
a
matter
of
debate,
in
vitro
studies
have
revealed
mechanisms
induced
by
endosulfan
that
are
implicated
in
tumor
development
and
progression
in
testes,
breast
and
liver
(
Hardell
and
Eriksson,
0300-483X/$–seefrontmatter © 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tox.2012.05.008
Table1
PrimaryantibodiesusedforWesternblotandimmunofluorescencestaining.
Antigen Phosphorylationsite Source/type Manufacturer DilutionWB DilutionIF
ROCK1 RabbitmAba Epitomics 1:2000 1:200
pFAK Tyr925 RabbitpAbb Cellsignaling 1:2000
FAK RabbitpAb Cellsignaling 1:2000
Vimentin RabbitmAb Epitomics 1:200
TCF1/LEF-1a RabbitmAb Epitomics 1:10,000
Gapdh RabbitmAb Cellsignaling 1:15,000
E-cadherin RabbitmAb Epitomics 1:5000
-Catenin RabbitmAb SantaCruz 1:2000 1:400
Fibronectin RabbitmAb Epitomics 1:1000 1:100
Snail RabbitpAb SantaCruz 1:200
Slug RabbitmAb Cellsignaling 1:300
amAb,monoclonalantibody. b pAb,polyclonalantibody.
2003;
Høyer
et
al.,
2002;
Fransson-Steen
et
al.,
1992;
Perez-Carreon
et
al.,
2009
).
An
example
of
such
a
mechanism
is
the
disruption
of
gap
junctions
(
Ruch
et
al.,
1990
).
Such
gap
junctional
intercellu-lar
communication
uncouplers
are
responsible
for
many
cellular
dysfunctions
and
multiple
diseases
including
cancer
(
Ehrlich
et
al.,
2006;
Trosko,
2011
).
In
addition,
the
genotoxic
potential
of
endo-sulfan
has
been
described
in
HepG2
cells
(
Hashizume
et
al.,
2010;
Li
et
al.,
2011
)
and
exposure
to
sub-lethal
doses
of
endosulfan
and
its
metabolites
causes
DNA
damage
and
mutations
(
Antherieu
et
al.,
2007;
Bajpayee
et
al.,
2006
).
One
study
using
a
computa-tional
quantum
chemical
model
indicated
that
endosulfan
and
all
its
metabolites
have
carcinogenic
potential
(
Bedor
et
al.,
2010
).
However,
the
exact
molecular
mechanisms
underlying
the
toxic
and
carcinogenic
effects
of
endosulfan
in
liver
cells
remain
poorly
understood.
Epithelial
to
mesenchymal
transition
(EMT)
is
a
highly
coor-dinated
multistep
process
during
which
epithelial
cells
acquire
mesenchymal
fibroblast-like
properties.
It
has
been
implicated
in
a
variety
of
diseases
including
fibrosis
and
in
the
progression
of
carcinoma
(
Jou
and
Diehl,
2010
).
This
fundamental
program
also
plays
a
key
role
in
the
critical
phases
of
embryonic
development
and
in
adults
it
contributes
on
the
one
hand
to
physiological
pro-cesses
such
as
tissue
repair
and
on
the
other
pathological
conditions
(
Kalluri
and
Weinberg,
2009
).
The
morphological
changes
imposed
are
governed
by
multiple
molecular
mechanisms
including
the
loss
of
epithelial
proteins
such
as
E-cadherin
(
Cavallaro
and
Christofori,
2004;
Huber
et
al.,
2005
)
that
leads
to
the
reduction
in
cell-to-cell
junctions
(adherens
and
gap
junctions),
the
loss
of
cell
polarity,
and
the
gain
of
mesenchymal
markers
such
as
the
extracellular
matrix
compound
fibronectin,
the
intermediate
filament
protein
vimentin
and
S100a4
(
Peinado
et
al.,
2004;
Kim
et
al.,
2011;
Ogunwobi
and
Liu,
2011
).
All
these
steps
lead
to
the
acquisition
of
motile
and
invasive
properties
(
Matsuo
et
al.,
2009
).
E-cadherin
is
thought
to
be
downregulated
via
several
repres-sors
acting
either
indirectly
(e.g.
Twist,
Goosecoid)
or
directly
by
binding
to
and
repressing
the
E-cadherin
promoter
(e.g.
Snail,
Slug,
Zeb)
(
Peinado
et
al.,
2007;
Yang
and
Weinberg,
2008
).
E-cadherin
repression
is
frequently
accompanied
by
the
activation
of
the
-catenin/Wnt
signaling
cascade
(
Fransvea
et
al.,
2008;
van
Zijl
et
al.,
2009a,b
).
Indeed,
during
EMT,
the
cytoplasmic
stabilization
of
-catenin
occurring
via
WNT
activation
signaling
and
GSK3-

repression
(
Medici
et
al.,
2008
),
leads
to
the
formation
of
a
nuclear
complex
TCF/LEF/
-catenin
that
modulates
the
expression
of
tar-get
genes
such
as
fibronectin,
MMP7
or
S100a4
(
Iwai
et
al.,
2010
).
Communications
between
cells
and
the
new
extracellular
matrix
environment
are
mainly
mediated
by
the
activation
of
the
focal
adhesion
kinase-FAK-transduction
pathway
(
Chatzizacharias
et
al.,
2008;
Cicchini
et
al.,
2008
).
The
liver
is
the
most
sensitive
tissue
to
xenobiotic
exposure
and
constitutes
the
main
target
of
endosulfan
toxicity.
Thus,
we
investigated
the
cellular
and
molecular
effects
of
endosulfan
in
human
liver
cells
using
the
HepG2
cell
line.
These
cells
retain
many
cellular
functions
allowing
the
study
of
several
molecular
pro-cesses
among
which
cell
plasticity
and
anoikis
(
Roe
et
al.,
1993
).
In
this
study,
we
show
that
endosulfan
permits
a
transient
anoikis
induction
that
is
followed
by
recovery
of
anoikis
resistance.
This
process
occurs
in
parallel
to
partial
EMT
and
is
characterized
by
the
repression
of
E-cadherin
expression,
the
loss
of
adherent
junctions,
cytoskeleton
reorganization,
extra-cellular
matrix
(ECM)
reshuffle,
WNT/
-catenin
pathway
activation
and
the
gain
of
mesenchymal
markers
(vimentin,
fibronectin,
S100a4).
Despite
evidence
of
all
these
events,
the
cells
remain
unable
to
acquire
a
motile
phenotype.
Thus
the
present
study
shows
that
endosulfan
transiently
sensitizes
HepG2
cells
to
anoikis
before
allowing
a
later
recovery
of
resistance
to
this
process
that
coincides
with
the
induction
of
partial
EMT.
Such
events
could
account,
at
least
in
part,
for
the
carcinogenic
potential
of
endosulfan.
2. Materialsandmethods
2.1. Materials
ThehumanhepatocellularcarcinomacellsHepG2wereobtainedfromATCC (AmericanTypeCultureCollection,Manassas,VA).Dulbecco’smodifiedEagle’s medium (DMEM), fetal bovine serum (FBS), penicillin/streptomycinsolution, sodiumpyruvateandEagle’snon-essentialaminoacidswerefromBioWhittaker (CambrexCompany,Walkersville,USA).DMSO(dimethylsulfoxide)andchemicals werefromSigma–Aldrich(L’Isled’AbeauChesne,SaintQuentinFallavier,France). ProteinassaymaterialswerefromBio-Rad.Allfluorescencereagentswerefrom MolecularProbes(Eugene,OR).TheOCendosulfanwasfromChemService(West Chester,PA).Theantibodiesusedforwesternblottingandimmunodetection exper-imentswerefromCellSignaling,SantaCruzandEpitomics(Table1).Cellswere visualizedwithaNikonEclipseTE2000phasecontrastmicroscope.
2.2. Cellcultureanddrugtreatments
HepG2cellsweremaintainedinDMEMwith1%penicillin/streptomycin,1%non essentialaminoacids,sodiumpyruvate,and10%FBS,inhumidifiedatmosphereat 37◦Ccontaining95%O2and5%CO2.Afterwashingwithsterilephosphatebuffer
saline(PBS),cellsweredetachedbytrypsinization(trypsin/EDTA)andplatedata concentrationof0.5–1×106or0.4–0.5×105cells/mlin6-wellor24-wellplates,
respectively,dependingontheexperiment.Forallexperimentalconditions,FBS wasreducedto5%inDMEMmediumandcellsweretreatedfortheindicatedtime withendosulfanpreparedasdimethylsulfoxide(DMSO)stocksolution.Thefinal DMSOconcentrationwas0.25%(v/v).
2.3. Viabilitytest
Viablecells were determinedby measuring theconversion ofthe tetra-zoliumsaltMTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide, Sigma–Aldrich(St.Louis,MO)toformazan,aspreviouslydescribed(Fautreletal., 1991).Briefly,cellswereseededin96wellplatesandtreatedat50%confluency withaconcentrationof20Mendosulfanduring24,48and72h.Thecellswere thenincubatedwith0.5mg/mlMTTfor2hat37◦C.Thewater-insolubleformazan
crystalwasdissolvedbyadding100lDMSOtoeachwellandtheabsorbancewas determinedwithaspectrophotometerat550nm(MR7000,DynatechLaboratories, Inc.,USA).
2.4. Real-timecellularimpedance
ThexCELLigencesystemwasusedaccordingtothemanufacturers’instructions (RocheAppliedScience,Mannheim,Germany)andACEABiosciences(SanDiego, CA,USA).Briefly,1×104cellswereaddedtothe96-wellE-plates.Twenty-four
hourslater,cellsweretreatedwithdifferentconcentrationsofendosulfan(2,10, 20,50,100and200M)in5%FBSDMEMmedium.Real-timecellularimpedance wasmeasuredineachwell(cellindexvalues)andasignalwasobservedthroughthe integratedsoftware(RTCAAnalyzer).Eachcurveisrepresentativeofanexperiment performedintriplicate.Tocomparetheinfluenceoftheendosulfanonthecells,the normalizedcellindex(NCI)wasused,calculatedusingtheDMSOcontrolcondition.
2.5. Immunofluorescence
Cellswereseededonglassslidesin24-wellplates(4×105cells perwell)
andwereexposedtoendosulfanattheindicatedtimeandconcentration.For cytoskeletonandextra-cellularmatrixproteins(Rho-associatedkinase,vimentin andfibronectin)andtranscriptionfactors(Snail,SlugandHif-1␣),thecellswere fixedwith4%paraformaldehyde(PFA)andpermeabilizedin0.5%saponin.For struc-turalproteins(E-cadherinand-catenin),cellswerefixedin−20◦Cmethanoland
permeabilizedin0.25%triton.Inbothcases,slideswereincubatedwithantibodies (Table1)for1hatroomtemperature.AfterwashingwithPBS,goatanti-rabbitor anti-mouseIgGcoupledtoAlexaFluor®488or594(MolecularProbes,Eugen,OR)
wereaddedtoslides,beforeincubationinadarkroomfor1hatroom tempera-ture.Theactincytoskeletonwasobservedafterincubationwithphalloidincoupled withAlexaFluor488for5min,andthenucleusafterincubationwithDAPI1g/ml (2,6-diamidino-2-phenylindone)for10min.Slidesweremountedandsealedin ProLongGoldantifadereagent(Invitrogen).Observationswereperformedwithan invertedfluorescencemicroscope(Nikon)equippedwithaCDDcamera(ORCA-ER HamamatsuPhotonics).
2.6. Detectionoftheanoikisprocess
Thecellsweretreatedfor48or72hwith2,10,20,50and100Mofendosulfan. Then,theculturemediacontainingfloatingcellswerecollected,centrifuged,and pelletsreseededintonew6-wellplateswithDMEMsupplementedwith5%FBS.This protocolisdesignedtotesttheabilityofcellstoadhereintheabsenceofendosulfan. Eighteendayslater,theMTTviabilitytestwasperformedtoestimatethenumber ofnon-adheringcellsfromtheobservedpopulationofadheringcells.
Secondly,todeterminethelevelofanoikis,2×105cellswereculturedasa
sus-pensiononplatescoatedwith(poly(2-hydroxyethylmethacrylate)(poly-HEMA) (Sigma)fortheindicatedperiodsoftime.Apoptoticcellswerethenestimatedby enzymaticassayusingcaspase-3activityasamarkerofapoptosis.
2.7. Enzymaticassaysforcaspaseactivity
Caspase activity was assessed by measuring fluorophore (7-amido-4-trifluoromethylcoumarin (AFC)) release from caspase tetrapeptide substrate N-acetyl-Asp-Glu-Val-Asp (Ac-DEVD) for caspases-3-like activity. Briefly, cells grownin6-wellculturedisheswerescrapedintoice-coldhypotonicbuffer.Cells werethenlysedbybeingsubjectedtothreecyclesoffreezingandthawing.Protein concentrationsweredeterminedusingtheBCAProteinAssaykit(Pierce,Rockford, IL,USA)andequalamountsweremixedwithbufferB(312.5mMHEPES,pH7.5, 31.25%sucrose,0.3125%CHAPS,50Mofrelativesubstrateenzymes).The fluo-rometricassaydetectedtheshiftinAFCfluorescenceemissionfollowingcleavage fromtetrapeptide-AFC,asmeasuredinafluorometer(ex=390nm;em=530nm).
2.8. Westernblot
HepG2cellswerescrapedintohypotonicbuffer(20mMHEPES,pH7.5,10mM KCl,15mMMgCL2,0.25mMsucrose,1mMEDTA,1mMEGTA,1mMDTT,0.1mM
PMSF,10g/mlpepstatinA,10g/mlleupeptin,andthephosphataseinhibitory cocktailPhosphoSTOP,Roche).Theproteinconcentrationineachcelllysatewas quantifiedusingaBCAProteinAssayKit(Pierce),withbovineserumalbumin(BSA) usedasastandard.EqualproteinamountswereseparatedbySDS-polyacrylamide gelelectrophoresison10%gelsandweretransferredtoPVDFmembranes.The membraneswereimmunoblottedwithantibodies(Table1)for1hatroom tem-peratureorovernight at4◦C.Afterwashing,membraneswerethenincubated
withhorseradishperoxidase-conjugatedsecondaryantibodies(anti-mouseor anti-rabbitimmunoglobulinG;Promega,Madison,WI,USA)for1hatroomtemperature. Afterwashing,thesignalsweredetectedusingImmobilonWesternDetection Reagents(Millipore,Molsheim,France)andacquiredusingaCCDcamera (Chemi-Genius2,SynGene).
2.9. Reversetranscription-quantitativepolymerasechainreaction
TotalRNAwasisolatedusingacidphenolextraction.OnemicrogramoftotalRNA wasreversetranscribedusingakit(SuperScriptII;InvitrogenCorp.,Carlsbad, Cal-ifornia)followingthemanufacturer’sinstructions.QuantitativePCRanalysiswas
carriedoutwithLightCycler®480ProbesMaster(Roche),accordingtothe
man-ufacturer’sinstructions,togetherwithFAM-labeledhydrolysisprobesfromthe UniversalHumanProbeLibrarySet(Roche).Intron-spanningprimersweredesigned usingtheUniversalProbeLibraryAssayDesignCentersoftware.Calculationswere madeusinggapdhastheendogenouscontrolreferencegene.Folddifferencesingene expressionwerecalculatedusingtheLightCyclersoftware,takingintoaccountthe efficiencyofamplificationasdeterminedfromastandardcurveobtainedwiththe second-derivativemaximummethod.
2.10. Cellmigrationassay
At90%confluenceHepG2cellsweretrypsinedand3×106ofcellsperwell
wereaddedtoa6wellplate.After24hofincubation,theconfluenttissueformed wasscratched4timesperwellwithasterilepipettetip.Cellswerewashedtwice withPBSmediumbeforebeingtreatedwith0.25%DMSOor20Mendosulfanina DMEMmediumdepletedwith5%FBS.Imagesweretakenimmediately(0h)with aninvertedfluorescencemicroscope(Nikon)equippedwithaCDDcamera (ORCA-ERHamamatsuPhotonics)andNIS-ElementsAR2.30softwareat4×magnification. 24hlater,thecellmigrationprogresswasphotographedinthesameconditions andalldataweretreatedwiththeTScratchsoftwaretooldevelopedforautomated analysisofmonolayerwoundhealingassaysasdescribedbyGebäcketal.(2009).
2.11. Statisticalanalysis
Eachexperimentwasrepeatedatleastthreetimes.Datashownarean aver-age±standarddeviation(SD).Statisticalanalysisofinvitrostudieswasperformed usingaStudent’sttest.Levelsofprobabilityareindicatedas*P<0.05or**P<0.01.
3.
Results
3.1.
Morphological
changes
in
HepG2
cells
after
endosulfan
treatment
To
assess
the
effect
of
endosulfan
on
hepatoma
cells,
HepG2
cells
were
treated
with
increasing
concentrations
of
the
organochlorine
pesticide.
Viability
was
evaluated
by
MTT
dye
reduction
assay
after
24,
48
and
72
h
of
treatment
or
by
measuring
cellular
impedance
in
real-time.
As
shown
in
Fig.
1
,
endosulfan
decreased
HepG2
viabil-ity
in
a
dose-
and
time-dependent
manner.
Cytotoxicity
occurred
within
72
h
following
50
M
endosulfan
and
peaked
after
100
M
(IC
50=
68.9
M
after
72
h
treatment).
Results
obtained
with
the
xCELLigence
system
do
not
allow
the
determination
of
IC50
values.
Indeed,
this
technology
is
a
non-invasive
cytotoxicity
assay
that
is
based
on
the
measurement
of
impedance
in
real
time
(cell
index
values).
This
parameter
reflects
the
cellular
status
with
regards
to
such
as
cell
proliferation,
vari-ations
in
cell
membrane
integrity,
cytotoxicity,
cell
adhesion
and
spreading,
and
cell
migration.
Fig.
1
B
depicts
the
dynamic
changes
in
cell
index
(CI)
values
of
cells
after
exposure
to
different
concen-trations
of
endosulfan.
At
100
and
200
M,
endosulfan
lead
to
a
slight
decline
in
the
cell
survival
rate
but
failed
to
induce
100%
mortality.
At
lower
concentrations,
a
significant
increase
in
the
slope
corresponding
to
the
evolution
of
cell
index
over
time
was
observed
(
Fig.
1
B
and
C).
This
phenomenon
could
be
explained
by
morphological
changes
correlating
with
defects
in
cell
adhesion
or
migration.
In
the
control
condition
(DMSO),
HepG2
cells
had
an
epithelial
cell-like
morphology
with
a
characteristic
“cobblestone”
appear-ance
and
organized
cortical
pattern
of
F-actin
at
cell-to-cell
junctions
(
Fig.
1
D).
By
contrast,
HepG2
cells
exposed
to
20
M
endosulfan
treatment
exhibited
more
poorly
defined
intercellular
borders
and
were
more
often
spherical
and
individualized
(arrows).
These
modifications
correlated
with
the
reduced
organization
of
F-actin
into
a
cortical
pattern
at
cell-to-cell
junctions
(
Fig.
1
D,
arrowhead).
These
results
suggested
that
endosulfan
is
toxic
at
high
concen-trations
(>50
M)
but
that
lower
doses
could
nevertheless
induce
phenotypic
modifications.
Fig.1.EndosulfanmodifiestheHepG2cellmorphology.(A)HepG2cellviabilitywasassessedbyMTTtestafter24h,48hand72htreatmentwithincreasingconcentrations ofendosulfan(from0.1Mto200M).MTTresultsarepresentedasapercentageofviabilityoverDMSOtreatmentandeachvalueisthemean±S.D.ofthreeseparate experiments(note:means±S.D.,*P<0.05and**P<0.001whencomparedtoDMSO).(BandC)Cellswereseededonto96-wellE-platesandtreatedfor24hwithendosulfan (from2,10,20,50,100and200M).Cellimpedancewasmeasuredinreal-timeandcellindexnormalizedagainsttheDMSOcondition.Resultsaremeans±S.D.for triplicatesofoneexperimentandarerepresentativeofthreeindependentexperiments.(D)48hafter20Mendosulfantreatment,cellmorphologywasexaminedundera lightmicroscope.F-actinwasvisualizedbyAlexaFluor488-conjugatedphalloidinstaining(green)andfluorescencemicroscopy.(Forinterpretationofthereferencestocolor inthisfigurelegend,thereaderisreferredtothewebversionofthearticle.)
3.2.
Endosulfan
sensitizes
HepG2
cells
to
anoikis
Based
upon
our
morphological
findings,
we
sought
to
evalu-ate
whether
endosulfan
affected
cell
death
processes.
We
observed
that
nuclei
from
individualized
cells
after
exposure
to
20
and
50
M
endosulfan
were
condensed
and
fragmented,
as
shown
by
DAPI
staining
(
Fig.
2
A).
Moreover,
when
these
floating
cells
were
replated
in
tissue
culture
dishes
after
48
h
or
72
h
in
suspension,
they
dis-played
the
ability
to
attach
and
proliferate
(
Fig.
2
B).
Indeed,
18
days
after
having
replated
the
cells,
MTT
test
showed
that
the
floating
cells
obtained
after
endosulfan
treatment
(<50
M)
had
grown
in
a
time-
and
dose-dependent
manner.
However,
at
high
concentra-tions
(50
and
100
M),
the
number
of
living
floating
cells
seemed
to
be
too
low,
as
demonstrated
by
the
low
viability
rate
obtained
after
replating
(
Fig.
2
C).
Hence,
these
results
suggest
that
the
process
leading
to
their
individualization
is
reversible.
We
firstly
inves-tigated
the
effect
of
endosulfan
on
capase-3
activity
levels
(as
a
marker
of
apoptosis)
in
both
the
adherent
and
the
floating
cells
after
48
h
endosulfan
treatment
(
Fig.
3
A).
The
caspase-3
activity
remained
stable
in
the
HepG2
cells
of
the
monolayer,
whereas
it
increased
significantly
in
the
floating
cells
at
concentrations
≥10
M
(400–450%
of
activation/DMSO).
Secondly,
to
investigate
whether
prevention
of
attachment
was
the
main
cause
of
cell
death
induced
by
endosulfan,
HepG2
cells
were
cultured
on
poly-HEMA
(+poly-HEMA),
which
prevented
cells
from
attaching.
The
cells
responded
with
a
significant
induction
of
caspase-3
activity
when
exposed
to
20
or
50
M
endosulfan
for
4
h
and
24
h,
but
not
with
DMSO
or
when
kept
under
adherent
conditions
(
−poly-HEMA).
However,
this
increased
caspase-3
activity
seemed
to
be
transient,
as
it
was
not
observed
after
48
h
treatment
(data
not
shown).
Moreover,
after
24
h
endosulfan
treatment
(20
M)
we
observed
an
induction
of
the
gene
encoding
X-linked
inhibitor
of
apoptosis
(XIAP),
the
endogenous
caspase-3
and
-7
inhibitor.
Thus,
endosulfan
initially
sensitized
the
HepG2
cells
to
anoikis
yet
per-mitted
the
restoration
of
intrinsic
anoikis
resistance
later
on.
Taken
together,
these
findings
demonstrate
that
endosulfan
may
induce
anoikis
in
HepG2
cells,
probably
via
disruption
of
cell
attachment.
3.3.
Endosulfan
induces
changes
in
the
cytoskeleton
and
ECM
associated
with
the
activation
of
FAK
signaling
and
potential
destabilization
of
adherent
junctions
We
therefore
investigated
whether
endosulfan
can
perturb
cel-lular
adhesion
by
studying
changes
in
the
expression
of
proteins
implicated
in
the
modification
of
the
actin
cytoskeleton
of
HepG2
cells
exposed
to
endosulfan.
We
examined
the
localization
and
expression
of
the
Rho-associated
kinase
(ROCK1),
a
well-known
effector
of
the
RhoGTPAses
implicated
in
lamelliopodia
and
stress
fiber
formation
during
cytoskeleton
remodeling
(
Schaller,
2010
).
Fig.2.Effectsofendosulfanonadherence,viabilityandnuclearcondensationofHepG2cells.(A)NuclearmorphologywasobservedbyDAPIstainingafter24hand48h ofDMSO(0.25%)orendosulfantreatments(20and50M).Arrowsindicateapoptoticnucleifromnon-adherentcellsafterendosulfantreatment.Theresultsshownare representativeofthreeseparateexperiments.(BandC)Floatingcellsobtainedafter48hor72hof20Mendosulfanexposurewerereplatedin6-wellplatesasdescribed inthematerialandmethodssection.18daysafterplating,picturesweretaken(B)andcellviabilitywasmeasuredbyMTTtest(C).
As
demonstrated
by
immunofluorescence
(
Fig.
4
A),
the
cellular
localization
of
ROCK1
was
modified
after
48
h
of
treatment
with
20
M
endosulfan.
Indeed,
in
the
control
condition
(DMSO),
ROCK1
was
mainly
colocalized
with
actin
leading
to
a
cortical
pattern
at
cell-to-cell
junctions,
whereas
in
the
presence
of
endosulfan,
this
protein
was
distributed
uniquely
throughout
the
cytoplasm.
We
also
examined
the
protein
levels
of
ROCK1
by
western
blot
analy-sis.
After
endosulfan
treatment,
the
expression
of
this
protein
was
up-regulated
in
a
time-dependent
manner
(
Fig.
4
B).
Focal
adhesion
kinase
(FAK)
signaling
was
also
implicated
in
the
control
of
the
actin
cytoskeleton
organization.
Indeed,
FAK
medi-ates
cells
motility
and
adhesion
turnover
through
regulation
of
the
RhoGTPases
and
ROCK
activity
(
Riento
and
Ridley,
2003
).
We
found
that
endosulfan
increased
the
phosphorylation
of
FAK
at
Tyr
925
in
a
time-dependent
manner
as
soon
as
8
h
of
endosulfan
treatment
(
Fig.
4
C).
The
phosphorylation
of
the
Tyr-925
residue
leads
to
cell
migration
and
cell
protrusion
(
Deramaudt
et
al.,
2011
).
Extracellular
matrix
(ECM)
remodeling
is
a
major
pheno-typic
modification
observed
during
cell
motility
and
migration.
Fibronectin,
a
high-molecular
weight
glycoprotein,
serves
as
a
scaf-fold
for
the
fibrillar
ECM.
We
therefore
analyzed
fibronectin
levels
by
western
blotting
and
immunofluorescence
studies.
Immunoflu-orescence
analysis
showed
an
important
increase
of
fibronectin
in
the
cytoplasm
of
HepG2
cells
after
48
h
of
endosulfan
treatment
(
Fig.
4
D).
These
findings
are
consistent
with
the
up-regulation
of
fibronectin
detected
by
western
blot
(
Fig.
4
E).
It
is
worth
noting
that
the
protein
level
of
fibronectin
was
increased
after
30
min
of
endosulfan
treatment
and
peaked
after
about
60
min.
Moreover,
fibronectin
was
deposited
as
fibrils
(
Fig.
4
F)
in
the
extracellular
compartment,
48
h
after
endosulfan
exposure.
Interestingly,
endo-sulfan
did
not
increase
the
gene
expression
of
Itga5,
corresponding
to
the
␣5
chain
of
the
fibronectin
receptor
integrin
␣51
(data
not
shown).
This
integrin
is
the
main
receptor
of
fibronectin
allowing
cells
to
migrate.
Together,
these
results
show
that
endosulfan
induced
cytoskele-ton
remodeling
and
a
modification
of
the
ECM
composition.
3.4.
Loss
of
adherent
junctions
through
E-cadherin
repression:
possible
involvement
of
the
repressors
Snail
and
Slug
To
further
confirm
the
effect
of
endosulfan
on
cell
adhesion,
we
investigated
E-cadherin
gene
and
protein
expression
levels
in
HepG2
cells
exposed
to
endosulfan.
Endosulfan
significantly
decreased
the
protein
levels
of
E-cadherin
after
24
h
and
72
h
of
endosulfan
treatment
(
Fig.
5
A).
Moreover,
E-cadherin
mRNA
expression
was
decreased
by
approximately
50%
after
48
h
endo-sulfan
treatment
(
Fig.
5
B).
These
data
suggested
that
E-cadherin
expression
is
repressed
in
HepG2
cells
at
the
transcriptional
level
after
endosulfan
exposure.
Hence,
we
sought
to
determine
the
expression
and
the
localization
of
Snail
and
Slug,
two
direct
Fig.3. Inductionofanoikisbyendosulfantreatment.(AandB)HepG2wereexposedtoincreasedconcentrationsofendosulfan(from2to50M)for48h.Caspase-3activities wereassayedasdescribedintheexperimentalproceduresineitheradherentcellsofthecellmonolayerorinfloatingcells(non-adherentcells)(A).(B)HepG2cellswere culturedovernighteitheradherent(−poly-HEMA)orsuspendedonpoly-HEMA-coatedplates(+poly-HEMA)andreceivedendosulfan(20or50M)for4,24or48hbefore measuringcaspase-3activity.(AandB)ResultsareexpressedasapercentageofthevalueobtainedforDMSO-treatedcells,setat100%.Dataaremeans±SDforfourseparate experiments.Caspase-3activitieswereassayedasdescribedintheexperimentalprocedures.(C)XIAPmRNAlevelwasassessedbyreal-timeRT-PCRafter4,24and48hof 20Mendosulfantreatment.RelativemRNAexpressionlevels(normalizedwithrespecttogapdh)weredeterminedandmRNAlevelsinDMSO-treatedcellsweresetto1. Errorbarsindicatethemeans±SEMoftriplicatedeterminationsfromthreeindependentexperiments.Note:*P>05and**P>0.01.
repressors
of
E-cadherin
transcription.
We
found
that
endosulfan
increased
Slug
and
Snail
mRNA
after
24
h
and
48
h,
respectively
(
Fig.
5
C).
Concomitantly,
we
observed
a
nuclear
translocation
of
these
two
factors
into
the
nucleus
of
HepG2
cells
treated
with
endosulfan
(
Fig.
5
D).
Since
a
decrease
in
levels
of
E-cadherin
is
a
widely
accepted
characteristic
associated
with
EMT,
we
wondered
whether
endo-sulfan
could
activate
this
process.
3.5.
Endosulfan
induces
stabilization
and
nuclear
translocation
of
ˇ-catenin
During
EMT,
loss
of
E-cadherin-mediated
cell–cell
adhesion
destabilizes
the
-catenin
interaction
with
actin
cytoskeleton
and
can
promote
its
activation.
Accordingly,
we
sought
to
evaluate
whether
endosulfan
could
act
on
the
-catenin
signaling
pathway
(Clevers,
2006).
After
endosulfan
treatment,
the
protein
level
of
-catenin
increased
in
a
time-dependent
manner
(
Fig.
6
A)
without
significant
modulation
of
its
gene
expression
(
Fig.
6
B).
This
protein
stabiliza-tion
occurred
together
with
the
nuclear
translocation
of
-catenin
after
48
h
endosulfan
treatment,
as
shown
by
immunofluorescence
imaging
in
Fig.
6
C.
Once
in
the
nucleus,
-catenin
binds
with
transcription
factors,
such
as
LEF/TCF,
to
activate
different
targets
genes,
such
as
LEF1,
TCF-1,
MMP-7
or
CCND1
(cyclin-D1)
(
Roose
and
Clevers,
1999;
Wielenga
et
al.,
1999;
Brabletz
et
al.,
1999;
Tetsu
and
McCormick,
1999
).
Endosulfan
increased
in
a
time-dependent
manner
the
protein
level
of
LEF1/TCF-1
when
compared
to
the
DMSO
condition
(
Fig.
6
D).
Moreover,
we
found
that
endosulfan
up-regulated
both
MMP-7
and
CCND1
at
the
mRNA
level
after
48
h
treatment.
These
results
suggest
that
endosulfan
activates
the
WNT/
-catenin
signaling
pathway.
3.6.
Endosulfan
induces
gain
of
the
mesenchymal
markers
S100a4
and
vimentin
but
does
not
confer
the
ability
to
migrate
EMT
is
characterized
by
loss
of
epithelial
and
gain
of
mes-enchymal
markers.
Consequently,
we
investigated
the
effect
of
endosulfan
on
the
expression
of
S100a4,
a
marker
of
hepatocytes
undergoing
EMT,
and
vimentin,
an
intermediate
filament
protein
normally
found
in
cells
of
mesenchymal
origin.
As
demonstrated
by
real-time
PCR,
endosulfan
increased
tran-siently
S100a4
gene
expression,
as
soon
as
4
h
and
peaked
after
24
h
endosulfan
treatment
(
Fig.
7
A).
Moreover,
we
observed
increased
amounts
of
vimentin
in
the
cytoplasm
of
HepG2
cells
exposed
to
endosulfan
(
Fig.
7
B).
These
results
would
suggest
that
endosulfan
leads
to
a
con-version
of
HepG2
cells
to
a
mesenchymal
phenotype
with
the
capacity
to
migrate.
Yet,
wound-healing
assays
revealed
no
Fig.4. EndosulfanperturbsthecytoskeletoninHepG2cellsandtheextracellularmatrix.(A,DandF)CellsweregrownoncoverslipsandtreatedeitherwithDMSO(0.25%) or20Mendosulfan.48hlater,cellswerefixedandprocessedforindirectimmunofluorescenceforthedetectionofactin(AandD,green),ROCK1(A,red)andfibronectin (DandF,red).Colocalizationwasindicatedbytheyellowcolorinthemergedimage.Representativeofthreeseparateexperiments.(B,CandE)HepG2cellsweretreated with20Mendosulfan.Attheindicatedtime,cellswerelysedandproteinlevelsoffibronectine(E),phospho-Fak,Fak(C)andRock1(B)analyzedbywesternblotting.As acontrol,thesamemembraneswerealsoprobedwithanantibodydirectedagainstGapdh.Thewesternresultsarerepresentativeofthreeindependentrepeatsforeach experiment.(Forinterpretationofthereferencestocolorinthisfigurelegend,thereaderisreferredtothewebversionofthearticle.)
significant
increase
in
wound
closure
after
24
h
or
48
h
endosulfan
treatment
when
compared
to
the
DMSO
condition
(
Fig.
7
C).
Thus
although
endosulfan
is
able
to
induce
major
changes
with
regards
to
the
expression
of
epithelial
characteristics,
it
does
not
confer
migration
properties
and
therefore
appears
to
induce
a
par-tial
or
incomplete
EMT
in
HepG2
cells.
4.
Discussion
In
this
study,
we
have
shown
that
endosulfan
transiently
sen-sitizes
HepG2
cells
to
anoikis,
disrupts
the
epithelial
phenotype
of
HepG2
and
induces
a
partial
EMT
process.
Anoikis
is
defined
as
cell
death
induced
by
inappropriate
or
loss
of
cell
adhesion
(
Frisch
and
Francis,
1994;
Meredith
et
al.,
1993
).
HepG2
carcinoma
cells
are
resistant
to
anoikis
but
after
endosul-fan
treatment
we
observed
an
induction
of
caspase-3
activity
in
poly-HEMA
culture
within
the
first
24
h.
After
that,
cells
recov-ered
their
resistance
to
anoikis
as
shown
by
increased
XIAP
mRNA
levels.
This
endogenous
caspase
inhibitor
protein
has
been
shown
to
contribute
to
the
anoikis
resistance
of
various
carcinoma
cells
(
Berezovskaya
et
al.,
2005;
Liu
et
al.,
2006
).
Hence,
overexpression
of
XIAP
in
cancer
cells
is
linked
to
their
increased
resistance
to
apo-ptosis
and
the
expression
level
reflects
apoptosis
sensitivity
(
Shi
et
al.,
2008
).
Acquisition
of
anoikis
resistance
is
a
likely
prerequisite
for
tumor
cells
to
successfully
metastasize
to
distant
sites
(
Liotta
and
Kohn,
2004
).
The
loss
of
cellular
contact
with
the
basement
membrane
constitutes
the
starting
point
of
the
onset
of
anoikis.
This
cell–ECM
interaction
is
monitored
by
integrins
that
bind
to
diverse
ECM
molecules
and
respond
by
triggering
an
intracellular
cascade
via
SRC
and
FAK
kinases.
We
found
that
endosulfan
increases
the
phosphorylation
of
FAK
at
the
Tyr
925
residue,
an
event
likely
to
contribute
toward
the
activation
of
downstream
signaling
events
including
the
Rho
GTPase
pathway.
This
sustained
phosphorylation
of
FAK
after
endosulfan
exposure
correlated
with
the
overexpres-sion
and
relocalization
of
the
cytoskeletal
RhoGTPase
proteins
(data
not
shown)
and
their
main
effector
ROCK1.
This
cascade
has
been
linked
to
cancer
cell
migration/invasion
during
metastasis
via
the
control
of
cytoskeletal
remodeling
(
Yamazaki
et
al.,
2005
).
FAK
is
a
major
protein
of
the
focal
adhesion
complex
that
integrates
signals
from
growth
factors
and
integrins
to
control
ECM
inter-actions,
migration
and
invasion
(
Mitra
et
al.,
2005
).
This
family
of
kinases
was
linked
to
tumor
invasiveness
and
their
expres-sion
and/or
activation
is
frequently
associated
with
metastatic
tumors
(
Gabarra-Niecko
et
al.,
2003;
McLean
et
al.,
2005
).
Notably,
in
parallel
to
FAK
phosphorylation,
we
found
an
up-regulation
of
fibronectin,
a
major
protein
that
serves
as
a
scaffold
for
the
fibrillar
ECM.
All
these
events
occurred
in
parallel
to
a
rearrange-ment
of
the
F-actin
cytoskeleton
with
formation
of
stress
fibers
in
Fig.5.EndosulfanrepressesE-cadherinexpression.(AandC)E-cadherin(CDH1),SnailandSlugmRNAlevelswereassessedbyreal-timeRT-PCRafter4,24and48hof20M endosulfantreatment.RelativemRNAexpressionlevels(normalizedwithrespecttogapdh)weredeterminedandmRNAlevelsinDMSO-treatedcellsweresetto1.Error barsindicatethemeans±SEMoftriplicatedeterminationsfromthreeindependentexperiments.(B)Attheindicatedtime,cellswerelysedandE-cadherinproteinlevels wereassessedbywesternblotting.(B)BandintensitieswereassessedbydensitometryafterimageacquisitionwithaCCDcameraandtheresultsarepresentedastheratioof Gapdh-normalizedresultsfortreatedcellstothoseforDMSO-treatedcells(means±SDforthreeexperiments).(D)After48hexposure,thecellswerefixedandprocessedfor indirectimmunofluorescenceanalysisforthedetectionofslugorsnail(red)andvisualizationofnuclei(DAPI,blue).Theresultsshownarerepresentativeofthreeseparate experiments.*P<0.05and**P<0.01.(Forinterpretationofthereferencestocolorinthisfigurelegend,thereaderisreferredtothewebversionofthearticle.)
